Background Body mass index (BMI) is correlated negatively with subjective well-being and positively with depressive symptoms. Whether these associations reflect causal effects is unclear.

Methods We examined bidirectional, causal effects between BMI and mental health with Mendelian randomisation using summary-level data from published genome-wide association studies (BMI: n=339 224; subjective well-being: n=204 966; depressive symptoms: n=161 460). Genetic variants robustly related to the exposure variable acted as instrumental variable to estimate causal effects. We combined estimates of individual genetic variants with inverse-variance weighted meta-analysis, weighted median regression and MR-Egger regression.

Results There was evidence for a causal, increasing effect of BMI on depressive symptoms and suggestive evidence for a decreasing effect of BMI on subjective well-being. We found no evidence for causality in the other direction.

Conclusion This study provides support for a higher BMI causing poorer mental health. Further research should corroborate these findings and explore mechanisms underlying this potential causality.